The Saproxylic Database: an Emerging Overview of the Biological Diversity in Dead Wood

THE SAPROXYLIC DATABASE: AN EMERGING OVERVIEW OF THE BIOLOGICAL DIVERSITY IN DEAD WOOD

1 2 Jogeir N. STOKLAND & Evgeniy MEYKE

RÉSUMÉ. — La base de données saproxyliques : l’émergence d’un aperçu de la diversité biologique du bois mort. — Les pays nordiques (Norvège, Suède, Finlande et Danemark) développent une base de données commune, en anglais, avec des informations spéciﬁ ques sur l’écologie des espèces saproxyliques. La base de données contient maintenant des informations sur environ 6000 espèces et ce nombre devrait avoisiner 7000 dans un proche avenir. Les principaux groupes d’organismes sont les Ascomycètes, les Basidiomycètes, les Coléoptères, les Diptères et les Hyménoptères. Le présent article décrit la structure de la base de données, l’information qui y est entrée pour chaque espèce et les différentes qualités de bois mort auxquelles les espèces sont associées. De plus il montre comment la base de données peut être utilisée et visualisée pour différents usages; quelques exemples de sorties de données sont fournis. Enﬁ n l’article décrit les utilisations potentielles, la maintenance et les plans de développement de la base de données dont son accessibilité sur Internet. Mots-clés: Base de données, pays nordiques, espèces saproxyliques.

SUMMARY. — The Nordic countries Norway, Sweden, Finland and Denmark are developing a shared English-written database with species-speciﬁ c information about the ecology of saproxylic species. The database now contains information for about 6000 species and the number is expected to approach 7000 species in the near future. The main organism groups are Ascomycetes, Basidiomycetes, Coleoptera, Diptera and Hymenoptera. This paper describes database structure, information that is entered for each species and different qualities of dead wood that the species are associated to. Furthermore, the paper describes how the database can be searched and viewed for different uses, and it gives some examples of data output. Finally, the paper describes potential uses, maintenance, and development plans including making the database accessible on Internet. Keywords: Database, Nordic countries, sproxylic species.

Dead wood is a key feature in natural forest ecosystems. Studies from unmanaged boreal forests indicate that dead wood makes up 20-30% of the total timber biomass (Syrjanen et al., 1994; Sippola et al., 1998), and up to 70% in recently disturbed areas (Krankina & Harmon, 1995). The amount of dead wood in North-European natural boreal forest varies between 60 and 80 m3/ ha, depending on latitude and altitude (Siitonen, 2001; Hahn & Christensen, 2004). In Central European temperate forest and mountain mixed forests, the amount of dead wood in natural forest conditions varies between 130 and 210 m3/ha (Hahn & Christensen, 2004). These amounts of dead wood represent forest conditions that have prevailed on our planet for many millions of years. During this period a large number of species have evolved to utilize this historically abundant resource base. These organisms are either directly involved in the degradation of dead wood, or they interact with the decomposers through fungivory, predation or parasitism.

1 University of Oslo, Natural History Museum. P.O. Box 1172. Blindern. N-0318 Oslo, Norway. E-mail: [email protected] 2 University of Helsinki, Finland. E-mail: evgeniy.meyke@helsinki.ﬁ

Rev. Écol. (Terre Vie), suppt. 10, 2008.

– 37 – In the Nordic countries there is a strong tradition to study the biological diversity in dead wood, rooted in classic works on beetles (Saalas, 1917; Palm, 1951, 1959) and fungi (Eriksson, 1958; Eriksson & Strid, 1969). Recently, such publications and unpublished data have been compiled in a database on the ecology of species living in dead wood (Dahlberg & Stokland, 2004). After the initial compilation a network of about 40 experts, mainly on different insect and fungus groups, have continued to update the database. Here, we present this work with emphasis on database structure and information content. Furthermore, we give some examples of output and suggest various uses of the database.

MATERIAL AND METHODS

BIOLOGICAL AND GEOGRAPHICAL SCOPE The aim of the saproxylic database is to compile species-speciﬁ c ecological information for all multi-cellular species that depend on decaying wood and to make this information available for different uses. The principal organism groups are Ascomycetes, Basidiomycetes, Coleoptera, Diptera and Hymenoptera. Furthermore, the database includes other insect groups (e.g. Hemiptera and Lepidoptera), additional invertebrate groups (e.g. mites and pseudoscorpions), vertebrates, mosses and lichens. Nearly all wood-inhabiting species live in terrestrial ecosystems (forests or parks), but some utilize wood submerged in freshwater or marine environments. Such aquatic species are also included in the database. The database and experts judgements indicate that the total species richness of wood-inhabiting species in the Nordic countries is close to 7000 (Tab. I). By the end of 2007, the number of species in the database was about 6000.

TABLE I Number of wood-associated species in different organism groups in Scandinavia1. The numbers are based on counting individual species or expert assessments Fungi Ascomycetes 750 Basidiomycetes 1270 Lichens 220 Insects Coleoptera (beetles) 1400 Diptera2 1000 - 1200 Hymenoptera2 800 - 1000 Other insects2 > 200 Other species groups Acarina (Mites)2 300 - 500 Nematodes2 > 100 Myxomycetes2 150 Mosses 90 Vertebrates 54 Total 6400 - 7000 (1) The numbers are based on data from Finland, Norway and Sweden. (2) The species numbers in these groups are expert assessments.

Denmark, Finland, Norway and Sweden represent the “target countries” for which the database provides species information. However, the database also includes information from surrounding countries (Fig. 1), especially for species that are rare or poorly studied within the Nordic countries. With time the database may be extended to cover a wider range of target countries.

DATABASE STRUCTURE The database is organized as a relational database structure with the following main tables: a) reference library, b) species occurrences, c) locality objects, d) wood attributes, e) species interactions, f) generic species information (Fig. 2). In a recent upgrading, the database was moved to a new software platform using Microsoft Visual Studio 2005 and C# as implementation tools and MySQL as database engine.

– 38 – Figure 1. — The geographical areas showing the four target countries and the most important countries representing the empirical basis for the saproxylic database.

Figure 2. — The main tables of the saproxylic database with lines showing how the tables are connected through linking ﬁ elds.

– 39 – The reference library is a table with a record for each information source used by the database. Publications in scientifi c journals, biological society journals and various reports represent the most important information sources. The database also includes information from unpublished data sets (from research projects, master studies and PhD studies). Such data sets often become available after the data have been analysed and the intended publications have been made. Finally, the database contains information from museum collections, individual fi eld observations, and expert judgments based on many years of fi eld experience. The species occurrence table essentially contains information about a species that is recorded in a place, with further links to ecology information in the wood attribute and species interaction tables, and more detailed geographical information in the locality table. The wood attribute and species interaction tables contain all the ecological information in the database. The wood attribute table contains information that documents species associations to different wood attributes. The species interaction table documents relationships between species at higher trophic levels depending on dead wood through their saproxylic hosts, prey species or symbiotic species (Fig. 3). It is a rule to specify the geographical origin for each primary record. The geographical areas follow the defi nitions of Brummitt (2001), and form a global, hierarchical system where the basic unit typically corresponds to a country (Fig. 1).

Figure 3. — Number of beetle species in Norway and Sweden that prefer or occur on individual tree species or genera.

The generic species table contains one and only one record for each species in the database. These records contain taxonomic position (Kingdom, Order, Family, etc.) the standard nomenclature (valid Latin name), national status (known or not known from each target country), and degree of wood dependency (obligate or facultative saproxylic).

PRIMARY SPECIES RECORDS The species occurrence, wood attribute and species interaction tables contain the primary species records. Each species occurrence record has a RefID link to the information source stored in the reference library. The species name contains the standard name, whereas used name specifi es the name used by the information source (if different from the standard name, e.g. old information using outdated nomenclature). The comments fi eld is a free text paraphrasing the information from the information source and is translated into English when the source uses another language. Development stage specifi es which part of the life cycle is documented (larva, pupa, adult; for fungi: mycelium, fruit- body). Each species occurrence record is linked to a wood attribute record where the information is coded as qualitative or quantitative observations according to a large number of specifi c wood attributes (see below). A species occurrence record can also be linked to another species occurrence record to form a species interaction record. This record defi nes the functional role of the species with respect to the other (e.g. wood decomposer, detrivore, fungivore, predator, parasite, scavenger, epixylic).

– 40 – The data is entered into Excel fi les according to specifi c guidelines outlined in Stokland (2007). These fi les are subsequently uploaded to the database through an import procedure that performs numerous automated checks of correct name spelling, valid codes and referential integrity between different tables.

Qualitative and quantitative records The association of a species to different wood attributes or another species is documented by using qualitative or quantitative records (see Tab. IIIb). The following codes are used for qualitative data: — preference for a particular wood attribute (as indicated by the information source). — occurrence on at least one wood unit with the particular wood attribute condition. This is the most commonly used code, as no assumption is made concerning preference for the wood attribute in question. — irregular occurrence. This code is used for an observation on a particular wood attribute where one would normally not expect to ﬁ nd the species. Quantitative records are used when the information source states the number of dead wood units or live trees on which the species was observed. In these cases the actual number is entered in the record instead of the qualitative codes.

Wood attributes The wood attribute table contains information about individual species and their usage of a large number of wood attributes. All the attributes have been defi ned as specifi c categories and most of them are listed below. A complete list of attributes and their defi nitions is presented in Stokland (2007). The wood attributes are grouped into 7 major classes: (1) Host tree. Host tree is a hierarchical category where one normally enters a species name (genus and epithet). But one can also enter higher categories such as genus (Betula spp., Quercus spp., etc.), “coniferous” or “broad-leaved tree”. One can also enter lower categories like subspecies and hybrids between trees. (2) Decay stages and rot types. The decay stage refers to the local state of decay that the species actually utilizes, and not some average decay of the whole tree (see wood types, below). The database uses two categories for live wood and fi ve for dead wood: Live wood, Dying wood, Recently dead, Weakly decayed, Medium decayed, Very decayed, Almost decomposed. In addition there are two categories for rot type: Brown rot, White rot. (3) Dimension. The database uses 7 diameter intervals to specify occurrence on different dimensions: 0-1 cm, 1-5 cm, 5-10 cm, 10-20 cm, 20-40 cm, 40-80 cm, > 80 cm. For species utilizing the trunk, the DBH is the reference diameter. For species utilizing the branches or tree top, the local branch (tree top) is the reference diameter. (4) Mortality factor. The mortality factor primarily focuses on how the tree died rather than the specifi c mortality agent (which can be several ones operating in combination). The following categories are used: Died standing, Up- rooted, Broken mechanically, Broken by rot, Fire, Cut by man. (5) Microhabitats. Nearly all saproxylic species utilize some distinct microhabitat(s) formed by the tree or other saproxylic species. A microhabitat is a smaller section of a live or dead tree. The most important ones are: Bark interior, Under bark, Wood surface, Wood interior, Insect galleries, Sap run, Rot hole, Tree cavity, Fire-charred wood. (6) Wood type. The wood type refers to the whole tree or unit of dead wood where a species is found. Several microhabitats can be present in a particular wood type, i.e. a live tree can contain a dead heartwood section or a tree cavity with wood in different stages of decay. The following categories of wood types are used: Live tree, Over-mature tree, Standing dead tree, Lying trunk or branch, Stump, Bolt, Logging residual, Building, Other modifi ed wood. (7) Surrounding environment. This class of wood attributes refers to the environment surrounding the wood unit where a species occurs. The following categories are used: Sun-exposed, Semi-shady, Shady, Buried in soil, Freshwater, Marine water.

Species interactions Perhaps more than a thousand saproxylic species in the Nordic countries depend upon woody material by feeding on other species that get their nutrition from woody material. Such species interactions can often be very speciﬁ c and the database is designed to capture these relationships using the species interaction table. Essentially the table has three elements: species A (on the “host” side), species B (the “associated” species), and the interaction form (fungus- fungivore, prey-predator, host-parasite, etc.). Furthermore, one can specify attributes such as the part of the life cycle being involved in the interaction both on the host and the associated side. Just like the wood attribute records, the species interaction records can be qualitative or quantitative.

AGGREGATION, SELECTION AND OUTPUT OF DATA

Aggregation levels The database contains primary species records with a high degree of ecological and geographical resolution. The information about each species is typically scattered across several records originating from different information sources. This resolution is often too detailed for various uses and the database is organized to present the information using four aggregation levels with decreasing amount of detail (Tab. II, Fig. 4). One can choose the aggregation level by specifying alternative views of the primary species records. Each view performs speciﬁ c pivot transformations of the data and presents the selected primary records according to the aggregation level. In the results we present some examples to illustrate these aggregation levels.

– 41 – The saproxylic database Users

Biodiversity statistics Forest management Aggregated species record Nature conservation Information dissemination Standard species record

Primary species records Research

Reference library Feedback - errors - missing info - new data s Data - type of use sets Publication Information sources Museum ions coll ecti Observat ons

Figure 4. — The Saproxylic database compiles data from a wide range of information sources. The data are aggregated and presented with different degree of detail for various uses. All users should provide feedback on shortcomings and usefulness of the database.

TABLE II The aggregation levels of species information from the saproxylic database

Aggregation level Characteristics Type of species information Biodiversity statistics Summary statistics of species numbers No individual species information. in different functional and ecological groups. Aggregated species record Each aggregated species record is Species classifi ed as specialist-generalist simplifi ed by joining similar wood with respect to wider wood attribute attributes into wider categories. categories. Standard species record All primary observation records for a Coded relationships (preference, species are combined into one summary occurrence, irregular occurrence) for record. Full resolution of wood attribute each wood attribute category. categories. Primary species record Each record refl ects what an individual Free text paraphrasing the information information source says about a source. Quantitative or qualitative data particular development stage of a for each wood attribute category. species.

Search criteria The aggregation level is one way to tune the output from the database. Another way is to select which parts to view. The search criteria can be taxonomical, i.e. one can specify which organism group, family, genus or species one wants information about. One can also specify the geographical scope, both in terms of target countries and empirical basis. When specifying the target country (or countries) one receives information about the species that are known from the country according to the national status in the generic species table. When specifying the empirical basis area one selects information from the primary species records that originate within this area. Thus, one can summarize the ecology of some species in a country based on knowledge from that country or a wider geographical area. Furthermore, one can select species that are obligate saproxylics, facultative saproxylics, or both.

Output formats The output from the database is presented on the screen for browsing and further reﬁ nement of the selection criteria. When the result is satisfying one can download the output as an Excel ﬁ le.

– 42 – RESULTS

PRIMARY RECORDS AND STANDARD SPECIES INFORMATION The primary species records represent the information basis of the database. Table III dis- plays a small subset of primary beetle records from the database. Table IIIa shows information that has been entered from the information source, and table IIIb shows how this information is coded into the wood attribute categories. In table IV, species information from table III is combined and transformed to standard species records. The main changes from the primary species records are the following: all comments, i.e. the text paraphrased from the information sources, are omitted; quantitative records are transformed to qualitative records; different preference levels (i.e. preference, occurrence, irregular occurrence) are combined, usually by giving “preference” and “irregular occurrence” priority over “occurrence”.

TABLE IIIA A selection of ﬁ elds from 6 primary species records in the species occurrence table

Ref Used Standard Geogr. Localiy ID name name DS Comments area ID The larvae develop under bark and in strongly decayed wood. It occurs in wood of broad-leaved trees. Searches of the species in hollow trees indicate that this is not a common Cetonia development place. It has been found in old, sun-exposed 53 aurata L manure heaps. Sweden The larvae develop in a particular kind of brown-rotted wood, typically in big standing or lying trunks that have been dead for 5-10 years. Often this kind of rot is found in narrow sections adjacent to cracks in the trunk. Occurs more Hylis often in Picea abies than Pinus sylvestris, also found in 53 cariniceps L several broad-leaved tree species. Sweden Larvae have been found in a Carpinus betulus branch on the ground. The larvae made galleries in the wood where a Hypocoelus reddish, somewhat moist rot had developed. The next year 303 cariniceps L Leiler reared the species from a dead Hedera helix stem. Sweden The species is almost exclusively confi ned to Pinus sylvestris, but in dry and very hot sites it probably also uses Picea abies. Larvae develop primarily in dead sun-exposed roots above ground, typically 5-10 cm thick. Often found in roots damaged by trampling and machinery; the natural habitat is probably trees/roots damaged by forest fi re. Larvae Buprestis have also been found in woody material on the ground in 53 octoguttata L sunny positions. Sweden The species primarily utilizes Populus tremula and different introduced Populus species. In northern Sweden the species also uses different Salix species. Larvae develop in 0.5 to 1 cm thin branches or stems of live trees (often ca 1 m high). The larvae fi rst live in the cambium zone and later Saperda enter the centre of the stem that develops a swollen gall 53 populnea L formation around the attack site. Sweden Larvae found in galls from 947 shoots of live Populus Saperda tremula (607), Salix caprea (188) and Salix phylicifolia 123 populnea L (152) in a study of parasitism and predation from enemies. Finland The species develops in dying or recently dead Betula pendula or Betula pubescens. The larvae develop in the Scolytus cambium zone of big trunks, thinner parts of the trunk in the 53 ratzeburgi L crown and big branches. Sweden RefID codes: 53 = Ehnström & Axelsson, 2002; 123 = Pulkkinen & Yang, 1984; 303 = Palm, 1959.

– 43 – TABLE IIIB The information from table IIIa coded into host tree and decay stages in the wood attribute table (p = preference, x = occurrence, i = irregular occurrence)

Host tree Decay stage spp. spp. Ref ID Species Pinus sylvestris Picea abies Betula tremula Populus Salix caprea Salix phylicifolia Salix Carpinus betulus helix Hedra trees Broad-leaved healthy Live, dying Live, dead Recently decayed Weakly Medium decayed decayed Very Almost decomposed 53 Cetonia aurata x xxx 53 Hylis cariniceps xp x p 303 Hypocoelus cariniceps 11 1 53 Buprestis octoguttata pi x 53 Saperda populnea p x p 123 Saperda populnea 607 188 152 947 53 Scolytus ratzeburgi p xx

TABLE IV The information from table IIIb converted to standard species records (p = preference, x = occurrence, i = irregular occurrence)

Host tree Decay stage spp. spp. Ref ID Species Pinus sylvestris Picea abies Betula tremula Populus Salix caprea Salix phylicifolia Salix Carpinus betulus helix Hedera trees Broad-leaved healthy Live, dying Live, dead Recently decayed Weakly Medium decayed decayed Very Almost decomposed 53 Cetonia aurata x xxx 53, 303 Hylis cariniceps xp xxx p 53 Buprestis octoguttata pi x 53, 123 Saperda populnea pxxx p 53 Scolytus ratzeburgi p xx

AGGREGATED SPECIES INFORMATION Many wood attributes have a hierarchical structure where some categories are ecologically similar to each other while others are very different. A species that only uses a narrow range of wood attributes within an attribute class is categorized as a specialist with respect to that class. All species are categorized as some kind of specialist, generalist or insufﬁ ciently known with respect to each of the seven wood attribute classes described in the methods section. Table V presents the species from table IV, but here they are classiﬁ ed as specialists or generalists relative to the wood attribute classes.

BIODIVERSITY STATISTICS The most general way to view the database content is to get summary statistics of species numbers that belong to some taxonomic group within a geographical area or are associated to

– 44 – TABLE V The records from table IV converted to aggregated species records

Species Host tree Decay stage Diameter class Cetonia aurata Broad-leaved specialist Late decay specialist Unknown Hylis cariniceps Conifer preference Mid decay specialist Medium-large preference Buprestis octoguttata Conifer specialist Early decay specialist Sma Saperda populnea Broad-leaved specialist Live specialist Very small specialist Scolytus ratzeburgi Broad-leaved specialist Early decay specialist Generalist

woody material in specifi c ways. Figure 3 illustrates this kind of biodiversity statistics. The fi gure shows the number of beetle species in Norway and Sweden (combined) that prefer or occur on individual tree species or genera. These numbers were derived from standard species records of 1257 beetle species in 2003. The result shows that among the species being observed on a tree species about 10% of them actually preferred this host tree. Another query based on the aggregated records of the same 1257 species revealed that 289 species (23%) preferred coniferous trees, 654 species (52%) preferred broadleaved trees, 138 species (11%) were generalists using both coniferous and broad-leaved trees, whereas 176 species (14%) were insuffi ciently known. In other words, a rather small proportion of the species preferred an individual tree species, but 75-90% of the species preferred either coniferous or broad-leaved trees. The uncertainty interval is caused by the insuffi ciently known species. The general picture presented above is quite reliable, although the exact numbers are outdated today. The reason is that the database is being updated with information about more species and some of the species that were in the database in 2003 have now been excluded.

DISCUSSION

In this paper we have presented the main features of the saproxylic database. We will now discuss some potential uses, routines of database updating, and further development (Fig. 4).

USES AND USER GROUPS

Scientifi c research The database has already been used as the basis for several scientifi c reports and papers (Dahlberg & Stokland, 2004; Stokland et al., 2004; Jonsson et al., 2005) and we foresee that researchers will be regular users of the database. One use is to extract information about species associations to different wood qualities and analyse responses of individual species and species richness in a predictive modelling context. Another use is to analyse species diversity patterns across different wood attributes and make review papers on the ecology of separate organism groups. A third, more indirect use, is to extract species classifi cations (obligate-facultative saproxylic, decay preference, host tree preferences, etc.) and use this information to sort species lists from fi eld studies. At least 5 scientifi c publications have already used the database for this purpose. Even if the database appears rich in content, it does have shortcomings. We therefore rec- ommend researchers who wish to use the database to contact the relevant working-group (see database maintenance, below) and check whether the database actually provides reliable data for the intended use. We have already discouraged specifi c uses because the results would be quite misleading based on the current database content.

– 45 – Forest management The database is relevant to adjust management practices directed towards the biodiversity in dead wood. By using the number of species associated to different dead wood attributes one can evaluate the cost-effectiveness of leaving different types of dead wood in the forest for biodiversity protection. Another use is to develop species-speciﬁ c conservation measures for management plans inside and surrounding localities of red-listed species. The database can also be used to deﬁ ne species guilds associated to different combinations of dead wood qualities (large-dimension and medium-decayed spruce logs, standing dead pine trees with loose bark, tree cavities in ageing oaks, etc.). Furthermore one can select some umbrella species representing each guild to give the biodiversity a “face”, and quantify the number of species underneath each umbrella.

Nature conservation The database can be a part of the knowledge base when supplementing existing networks of forest reserves. By quantifying the number of species being associated to different tree species one can assess to which extent different forest types are over-represented or under- represented with respect to cover the full range of species diversity in dead wood. When the species composition within established reserves is known one can use the database to assess favorable conservation status for the saproxylic species based on their substrate requirements and the actual availability of these substrate qualities.

Feedback on information content As the database becomes accessible for various uses, different users will spot errors, know missing information that ought to be in the database, and hopefully provide new data. We consider feedback from users to be instrumental for a successful development of the database (Fig. 4).

DATABASE UPDATING AND MAINTENANCE The database is developed and maintained by separate working groups established within the Nordic saproxylic network. These working groups carry out data entry and quality control for separate taxonomic groups. Most of the activity in the working groups has focused on data entry. Since the expert network was established in 2004, we have screened and digitized information from more than thousand papers, books, data sets, etc. This work will continue on a large scale at least until 2008, and hopefully we will manage to enter most of the available information from the Nor- dic countries. We have experienced that a high level of expertise is needed among the persons participating in the data entry process. One needs to correctly interpret what is stated by the information source, one needs expertise on the ecology of saproxylic species, and fi nally one must understand the wood attribute defi nitions in the database in order to translate the original information into these categories. We have also started to enhance the quality of the database content. One part of this work is to sort out exactly which species qualifi es as saproxylic. In this work we classify the species as “obligate saproxylic” - completely dependent upon dead wood for fulfi lling at least one part of the life cycle; “facultative saproxylic” - a large proportion (minimum 30-40% of known occurrences) utilize dead wood in a particular life cycle stage, but alternative resources is also used; “saproxylic” – either obligate or facultative, but existing data is too sparse to decide. The challenge is to sort out all the borderline cases among facultative species that partly use woody material (or other saproxylic species) and partly use other media or species as alternative resources. Another aspect of quality enhancement is to highlight or correct misleading/erroneous information from the information sources. A typical example is to update old nomenclature by entering the currently valid name in addition to the name used by the information source.

– 46 – ON-GOING DEVELOPMENT Until 2006, the database has been developed as a collection of separate fi les: an EndNote fi le for the reference library and a set of Excel fi les for the other tables illustrated in fi gure 2. This platform is unsatisfactory for several reasons and the database went through a major revision in 2006 and 2007. One change was to revise the internal database structure. The most important revision was to merge host tree associations and species interactions in a common structure. In the new version the wood from host trees is viewed as a species (which it in fact is) just like any other species. Then all records will be species-species interactions where one can specify any kind of ecological interaction (e.g. host tree – decomposer, fungus – fungivore, prey – predator, etc.) and various attributes specifying details about the species on either side of the relationship. We will also structure the wood attributes in a hierarchy that makes it straightforward to join detailed attribute categories into broader categories. Another change is to use a software platform that facilitates database access through Inter- net. From 2007 the database has become available on Internet at the address www.saproxylic. org. The Internet version is still under construction as we continue to develop new features to facilitate different uses of the database. In the near future we intend to introduce additional features both for data providers and database users. One feature is to add new fi elds to link species occurrences to specimens resid- ing in collections (private or public). Another feature is allowing data providers to attach special licences to previously unpublished data specifying that users need to contact the data provider if he or she wants to use the data for specifi c purposes (e.g. publish a report or scientifi c article). Still another feature is to allow users to express doubt concerning the validity of specifi c pieces of information in the database. In this way, the database can serve as interactive platform for cleaning up misleading or wrong information about saproxylic species.

ACKNOWLEDGEMENTS

We wish to express the warmest thanks to all colleagues and friends in the saproxylic network who have shared their knowledge and provided a lot of input and inspiration. Without this network and co-operation it would be impossible to develop the database so thoroughly. We also wish to thank the Swedish National Board of Forestry, Swedish Species Information Centre, Swedish University of Agricultural Sciences (Fortlöpande miljöanalys), Norwegian Species Information Centre, and the Ministry of Environment in Finland. Their funding has been instrumental for developing the database.

REFERENCES

BRUMMITT, R.K. (2001). — World geographical scheme for recording plant distribution. http://www.tdwg.org/geo2. htm, http://www.rbgkew.org.uk/gis/tdwg/ DAHLBERG, A. & STOKLAND, J.N. (2004). — Vedlevande arters krav på substrat – en sammanställning och analys av 3600 arter (Substrate requirements of wood-inhabiting species – a synthesis and analysis of 3600 species). Skogsstyrelsen, report 7, 2004. In Swedish with English summary. EHNSTRÖM, B. & AXELSSON, R. (2002). — Insektgnag i bark och ved [Insect galleries in bark and wood]. Uppsala, Artdatabanken. ERIKSSON, J. (1958). — Studies Göteborgs in the Heterobasidiomycetes and Homobasidiomycetes - Aphyllophorales of Muddus National Park in North Sweden. Symbolae Botanicae Upsalienses, 16: 3-172. ERIKSSON, J. & STRID, Å. (1969). — Studies in the Aphyllophorales (Basidiomycetes) of northern Finland. Annales Universitatis Turkuensis (A II), 40: 112-158. HAHN, K. & CHRISTENSEN, M. (2004). — Dead wood in European forest reserves – a reference for forest management. In: M. Marchetti (ed.) Monitoring and indicators of forest biodiversity in Europe – from ideas to operationality. EFI Proceedings, 51: 181-191. JONSSON, M., RANIUS, T., EKVALL, H., BOSTED, G., DAHLBERG, A., EHNSTRÖM, B., NORDÉN, B. & STOKLAND, J. N. (2005). — Cost-effectiveness of silvicultural measures to increase substrate availability for red-listed wood- living organisms in Norway spruce forests. Biol. Cons., 127: 443-462. KRANKINA, O.N. & HARMOND, M.E. (1995). — Dynamics of dead wood carbon pool in Northwestern Russian boreal forests. Water, Air and Pollution, 82: 227-238.

– 47 – PALM, T. (1951). — Die Holz- und Rinden-käfer der nordschwedischen Laubbaüme. Medd. 40 (2). Statens Skogforsk- ningsinstitut, Stockholm. PALM, T. (1959). — Die Holz- und Rinden-käfer der süd- und mittelschwedischen Laubbaüme. Opuscula Entomo- logica, Suppl. XVI. PULKKINEN, M. & YANG, Z.-Q. (1984). — The parasitoids and predators of Saperda populnea (Linnaeus) (Coleoptera, Cerambycidae) in Finland. Ann. Entom. Fennici, 50: 7-12. SAALAS, U. (1917). — Die Fichtenkäfer Findlands I. Ann. Acad. Sci. Fenn., ser. A., 8: 1-547. SIITONEN, J. (2001). — Forest management, coarse woody debris and saproxylic organisms: Fennoscandian boreal forests as an example. Ecol. Bull., 49:11-41. SIPPOLA, A.-L., SIITONEN, J. & KALLIO, R. (1998). — Amount and quality of coarse woody debris in natural and man- aged coniferous forests near the timberline in Finnish Lapland. Scand. J. For. Res., 13: 201-214. STOKLAND, J.N. (2007). — The saproxylic database - description and internal user guidelines. Report available at www. saproxylic.org. STOKLAND, J.N., TOMTER, S. & SÖDERBERG, U. (2004). — Development of dead wood indicators for biodiversity monitoring: experiences from Scandinavia. In: Marchetti, M. (ed.) Monitoring and indicators of forest biodiversity in Europe – from ideas to operationality. EFI Proceedings, 51: 207-226. SYRJANEN, K., KALLIOLA, R., PUOLASMAA, A. & MATTSON, J. (1994). — Landscape structure and forest dynamics in subcontinental Russian European taiga. Ann. Zool. Fenn., 31: 19-34.

– 48 –